👉 Join our Telegram group for updates and feedback
Model Overview — Wan2.2-Remix
Wan2.2-Remix is a research and creative model designed for producing short, imaginative video clips from textual prompts. It explores human figure dynamics, body movement expression, and scene consistency, delivering smoother motion and enhanced realism in character interactions. No additional LoRA setup is required.
- High-noise model: Lightx2v Wan2.2 Lightning dyno version (quantized to fp8)
- Low-noise model: Comfy-Org Wan 2.2 ComfyUI Repackaged (fp16, quantized to fp8)
- Enhancements: Blended with open-source LoRA resources and customized training data to enrich motion, body posture, and conceptual fidelity.
- Integration: Already bundled with
lightx2v, no extra setup required - Recommended CLIP: NSFW-Wan-UMT5-XXL
This release is currently beta. While functional, it is not yet perfect—feedback and testing are highly appreciated.
Changelog
v0.9
- High-noise model updated to Lightx2v dyno version, quantized to fp8
- Low-noise model updated to Comfy-Org fp16 version, quantized to fp8
- Improved character motion coherence
- Enhanced natural human limb articulation
v0.8
- Refined body motion balance for smoother expression
- Improved anatomical rendering for both male and female figures
- Adjusted walking animation weights for realism
- Enhanced visual clarity in intimate interaction scenarios
v0.7
- Introduced basic body dynamics simulation
- Added anatomical detail to human figures
- Improved facial rendering and expression consistency
- Added dynamic actions: walking, dancing, running
- Enhanced pose generation and overall gesture variety
Disclaimer
This model is released for research and creative purposes only. It must not be used for illegal, unethical, or harmful activities. The author assumes no responsibility for any misuse or consequences arising from its use.
模型概述 — Wan2.2-Remix
Wan2.2-Remix 是一款研究與創意模型，可根據文字提示生成 短片段的影像作品。 本模型著重於人體姿態、肢體動作表現力以及場景一致性， 讓人物動態更自然流暢，互動更加真實。 無需額外導入 LoRA。
- 高躁模型： Lightx2v Wan2.2 Lightning dyno 版本 （量化至 fp8）
- 低躁模型： Comfy-Org Wan 2.2 ComfyUI Repackaged （fp16，量化至 fp8）
- 增強： 融合開源 LoRA 資源與自訂訓練數據， 提升動作表現、姿態細節與概念一致性。
- 整合： 已內建
lightx2v，無需額外設置 - 推薦 CLIP： NSFW-Wan-UMT5-XXL
本版本目前為 測試階段，仍存在改進空間，歡迎回饋與測試。
更新日誌
v0.9
- 高躁模型更新為 Lightx2v dyno 版本，量化為 fp8
- 低躁模型更新為 Comfy-Org fp16 版本，量化為 fp8
- 改善人物動作的協調性
- 增強人體肢體的自然表現
v0.8
- 優化肢體動態平衡，使表現更自然
- 改善男女角色的解剖細節呈現
- 調整行走動作權重以提升真實感
- 增強互動場景中的細節與清晰度
v0.7
- 引入基礎的肢體動態模擬
- 增加人體解剖細節
- 改善臉部渲染與表情一致性
- 新增動作：行走、跳舞、奔跑
- 提升姿態生成與動作多樣性
免責聲明
本模型僅供研究與創意用途。 禁止用於任何非法、不道德或有害的行為。 作者不承擔任何因濫用而產生的後果。